Inspiration
As a leader of the Diversity, Equity, and Inclusion (DEI) project within our IT company, I’ve seen first-hand the barriers women face in the tech industry. The development teams often lack gender diversity, and this imbalance directly impacts innovation, productivity, and team creativity. Inspired by the ongoing need to close the gender gap in tech, I wanted to create a tool that not only provides insights into the problem but also makes these insights accessible to non-technical stakeholders. Everyone, from executives to team leads, needs to understand the data and to clearly see the disparities within the current system. I wanted to create something that makes the issue visually clear, actionable, and data-driven, enabling us to identify precise areas for improvement.
What it does
The Diversity Impact AI Tool empowers anyone—regardless of their technical background—to analyze gender gaps within our development teams. By leveraging AI, the tool allows users to choose from a set of predefined analysis options and generates dynamic, graphical data insights. Whether it’s the distribution of genders across roles, the correlation between gender and career progression, or the impact of gender on technical skills and job satisfaction, the tool transforms raw data into easy-to-understand visualizations. The backbone of the tool pulls in data through Gemini’s AI capabilities, generating both visualizations and code to present clear, actionable insights into gender diversity.
How we built it
We began by understanding the critical gender diversity metrics we wanted to track and analyze. Using Python, Pandas for data manipulation, and Seaborn for visualization, we constructed a user-friendly tool that integrates with Google's Gemini AI. The AI helps users automatically generate relevant datasets and code with minimal interaction. Once users input the dataset columns, the tool dynamically generates analysis options focused on common gender gap issues in tech, such as job role distribution, career progression, and employee satisfaction. From there, the tool automates the data processing and generates visualizations that can be easily shared across the organization.
Challenges we ran into
The biggest challenge was ensuring that the tool would be useful for non-engineers. While the concept was relatively straightforward for data scientists and engineers, translating complex processes into a simplified user experience for non-technical stakeholders required iterative testing and feedback. Another major obstacle was ensuring the accuracy of the data being processed. Datasets often contained missing entries, incorrect data types, and inconsistencies—requiring additional data cleaning logic and error handling features to be implemented. Integration with Gemini also posed minor hurdles, particularly with ensuring that the AI-generated code aligned perfectly with the graphs and insights we wanted to showcase.
Accomplishments that we're proud of
We’re proud that the Diversity Impact AI Tool effectively merges data-driven insights with accessibility. Non-engineers within our company can now confidently run gender gap analyses, explore visualized data, and receive code templates for further customization. This tool has democratized access to important gender diversity data, helping to foster a culture of inclusion and transparency. Seeing that this tool can empower leadership teams to easily identify gaps and take informed actions is a significant milestone for our DEI initiative.
What we learned
Through this project, we learned how critical user-centered design is for complex analytical tools. The primary audience for this tool—non-engineers—needed a high level of ease-of-use, and we learned that providing them with automation, while maintaining accuracy, was key to success. We also gained valuable insights into the importance of continuous iteration and user feedback; each round of testing helped refine both user interaction and the quality of the insights. Finally, we realized how essential clean data is. Without accurate input, the insights would be skewed, so we improved our capabilities around data validation and cleaning.
What's next for Diversity Impact AI tool
Next, we plan to incorporate data from more diverse sources, enriching the analyses with internal and external datasets to provide a more comprehensive view of team diversity. By leveraging a variety of data, we can offer deeper insights and powerful benchmarking for our company.
We also aim to enhance usability by turning the tool into a Web-based dashboard, allowing leadership and non-technical users to access real-time, interactive visualizations with ease. This will provide a more accessible, intuitive platform for exploring diversity metrics.
In addition, we plan to include AI-driven recommendations and real-time monitoring, driving actionable solutions to the diversity gaps identified. Ultimately, the goal is to broaden adoption across teams, embedding this tool as a key part of our diversity and inclusion strategy.
Log in or sign up for Devpost to join the conversation.